Thoughts about loudspeakers
Last edited: Dec. 3, 2019
Explanations of the 'boxy sound', extraordinary speaker designs, full-range drivers, two-way vs. three-way speakers.
The most important component of an audio system is the loudspeaker and the headphone. Modern audio players and amplifiers are so perfect that it makes no sense to talk about audible differences anymore. On the other hand the difference between loudspeakers and headphones is still huge. Most loudspeakers have a nonlinear frequency response (may come from driver resonances, uncompensated baffle diffraction, excessive cabinet edge diffraction, etc.) or other problems (poor off-axis responses). In addition to that the speaker's frequency response will vary according to room size, speaker placement, listening distance... Reflection from the wall behind the speaker and from the floor (or from the table for desktop speakers) is a common problem in room acoustics.
Explanations of the 'boxy sound' - what makes a loudspeaker's sound 'boxy':
- A loudspeaker without baffle step compensation will sound thin if set up far away from the rear wall. The loss depends on the baffle dimensions. (More about baffle step with links on the page of Rear Wall Reflection Simulator)
- Small PC speakers, desktop speakers have little or no bass and typically an overshoot (boost) between 200 Hz and 300 Hz. So in this case the reason for the boxy sound is the poor bass reproduction of the full-range driver, the small cabinet size and not the cabinet material and construction. (Although there is also a lack of baffle step compensation too.)
- There is some sound leakage from poorly constructed large cabinets (> 5 liters) with thin walls and no damping material. The sound leakage is greatest at mid frequencies.
Most extraordinary speaker designs have no any sonic advantage over the traditional two-way / three-way speaker design:
- Transient perfect (linear phase) speakers with time aligned drivers. The audibility of phase response and the impulse response are exaggerated by some speaker manufacturers. Our hearing becomes sensitive to phase (change of group delay) only in the lower frequency range, mainly below 200 Hz.
- The magic of point source: full range speakers, coaxial speakers are actually a compromise, not an advantage.
- Open baffles, omnidirectional speakers...
- Transmission line speakers.
- Speakers with spherical or spiral shell-like cabinet....
- Speaker cabinets made of stone, glass, marble or concrete... These materials are actually worse than plywood or MDF and they are not for speaker cabinet building.
- Bi-amplification and bi-cabling. Bi-cabling is utter nonsense and active crossovers are necessary only for bass cabinets.
- Ribbon tweeters, Air Motion Transformer tweeters, dome midranges. There is more potential in traditional dome tweeters and cone midranges.
Single full-range drivers (aka one-way speakers) are a compromise compared to multi-way speakers. The problem with full-range speakers is that no driver can reproduce the full audio range. Reproduction of bass frequencies requires large cone diameter with large cone excursion. Reproduction of the treble range needs small cone, and the cone excursion is not an important factor. Small full-range speakers (in the 2-3 inch diameter range) have good high frequency reproduction, but limited bass, and on the contrary large full-range speakers have more bass but they lose in the treble. For small desktop speakers, 3 inch full-range drivers are really good choice, because they represent a good balance between size and frequency response.
Two way vs. three way speakers. None of the two designs have any magical qualities compared to the other. Three way design make sense only if the woofer is larger than 6 inches. In this case three way configuration simplifies tweeter selection and crossover design.
The difference between microphone measurements and human hearing has been known for a long time and has been available in the form of measurements since the 1970s (HRTF measurements). This difference is important when interpreting room acoustic issues. For example, due to the directivity of the human ear, the effect of early reflections from a certain angle cannot be corrected with an EQ based on microphone measurements above 500 Hz.